lable at ScienceDirect

Journal of Archaeological Science 53 (2015) 550e558

Contents lists avai

Journal of Archaeological Science

journal homepage: http: / /www.elsevier .com/locate/ jas

Neutron activation analysis of 12,900-year-old stone artifacts confirms450e510þ km Clovis tool-stone acquisition at Paleo Crossing(33ME274), northeast Ohio, U.S.A.

Matthew T. Boulanger a, b, *, Briggs Buchanan c, Michael J. O'Brien b, Brian G. Redmond d,Michael D. Glascock a, Metin I. Eren b, d, *

a Archaeometry Laboratory, University of Missouri Research Reactor, Columbia, MO, 65211, USAb Department of Anthropology, University of Missouri, Columbia, MO, 65211, USAc Department of Anthropology, University of Tulsa, Tulsa, OK, 74104, USAd Department of Archaeology, Cleveland Museum of Natural History, Cleveland, OH, 44106-1767, USA

a r t i c l e i n f o

Article history:Received 30 August 2014Received in revised form1 November 2014Accepted 5 November 2014Available online 13 November 2014

Keywords:ClovisColonizationLong-distance resource acquisitionNeutron activation analysisNorth American Great LakesPeopling of North AmericaPaleo Crossing siteOhioWyandotte chert

* Corresponding authors. Department of AnthropoColumbia, MO, 65211, USA.

E-mail addresses: erenm@missouri.edu (M.I. Eren(M.T. Boulanger).

http://dx.doi.org/10.1016/j.jas.2014.11.0050305-4403/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

The archaeologically sudden appearance of Clovis artifacts (13,500e12,500 calibrated years ago) acrossPleistocene North America documents one of the broadest and most rapid expansions of any cultureknown from prehistory. One long-asserted hallmark of the Clovis culture and its rapid expansion is thelong-distance acquisition of “exotic” stone used for tool manufacture, given that this behavior would beconsistent with geographically widespread social contact and territorial permeability among mobilehunteregatherer populations. Here we present geochemical evidence acquired from neutron activationanalysis (NAA) of stone flaking debris from the Paleo Crossing site, a 12,900-year-old Clovis camp innortheastern Ohio. These data indicate that the majority stone raw material at Paleo Crossing originatesfrom the Wyandotte chert source area in Harrison County, Indiana, a straight-line distance of 450e510 km. Our analyses thus geochemically confirm an extreme stone-source-to-camp-site distance of aClovis site in eastern North America and thus provide strong inferential material evidence that the fastexpansion of the Clovis culture across the continent occurred as a result of a geographically widespreadhunteregatherer social network.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The archaeologically sudden appearance of Clovis artifacts acrossNorth America documents one of the broadest and most rapid ex-pansions of any culture known from prehistory (Goebel et al., 2008;Hamilton and Buchanan, 2007; Meltzer, 2009; O'Brien et al., 2014;Prasciunas and Surovell, in press; Smallwood and Jennings, inpress; Waters and Stafford, 2007). Current estimates suggest thatthe appearance of the Clovis culture occurred in the American Westca. 13,500e12,800 calibrated years before present (calBP) and in theEast ca.12,800e12,500 calBP (Ferring, 2001; Gingerich, 2011; Hayneset al., 1984, 2007; Holliday, 2000; Levine, 1990; Sanchez et al., 2014).Based on these dates, Clovis tools, including fluted lanceolate

logy, University of Missouri,

), boulangerm@missouri.edu

projectile points, spurred and notched endscrapers, prismatic blades,and bone and ivory rods, are among the earliest artifacts in NorthAmerica.

One long-asserted hallmark of the Clovis culture and its rapidexpansion is the long-distance acquisition of “exotic” stone used fortool manufacture (Meltzer, 2009; Bradley et al., 2010; Ellis, 2008;Frison and Bradley, 1999; Goodyear, 1989; Haynes, 2002; Kellyand Todd, 1988; Kilby, 2008; MacDonald, 1968; Speth et al., 2013;Stanford and Jodry, 1988; Witthoft, 1952). The question remainsas to whether this acquisition occurred “directly,” that is, by Clovisforagers obtaining stone from a source themselves (Ellis, 2011), or“indirectly,” meaning the stone went through one or more in-termediaries or groups between source and users (Hayden, 1982;Speth et al., 2013). With respect to direct acquisition, there is alsodebate as to whether stone procurement was “embedded”within aband's annual mobility round or whether specialized task groupsmade long-distance treks to the sources (Seeman,1994; Speth et al.,2013).

1 This date is an average of several selected AMS dates. The dated subsurfacefeature (B XVII F#1) yielded dates of 10,800 ± 185 14C YBP, 10,980 ± 110 14C YBP,11,060 ± 120 14C YBP, 12,000 ± 110 14C YBP, and 12,175 ± 115 14C YBP. These dates“define two cohesive but statistically distinct groups of radiocarbon ages; at12,150 ± 75 B.P. and 10,980 ± 75 B.P.” (Brose, 1994:65). Given that the older group ofdates seems too old for a Clovis site in the Great Lakes region (Brose, 1994), and theyounger group is consistent with Clovis dates elsewhere, Brose (1994:65) suggestedthe “most likely age of the [Clovis] occupation is reflected in the averaged date ofthe younger statistically distinct cluster of dates.”

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558 551

Whereas there exists archaeological equifinality for all the ma-terial correlates of these distinct possibilities (Bamforth, 2009;Burke, 2006; Meltzer, 1989; Speth et al., 2013), empirically vali-dating the long-distance acquisition of stone by Clovis foragersd-whatever the acquisition proceduredhas implications for the rapidexpansion of Clovis culture. Given the low density of Clovis sitesacross North America, and thus presumably low population den-sities, empirical validation of long-distance acquisition of stoneprovides a unique opportunity for inferring geographically wide-spread social contact and territorial permeability among mobilehunteregatherer populations (Bamforth, 2009; Burke, 2006; Ellis,1989; Kelly and Todd, 1988; Pearce, 2014; Pearce and Moutsiou,2014; Speth et al., 2013). For example, in terms of tool flaking pat-terns, there appears to have been a continent-wide standardizationof Clovis technology, perhaps reflecting Clovis knappers sharingtheir technical knowledge through direct transmission (Sholts et al.,2012). Sholts et al. (2012) suggest this transmission could haveoccurred at quarries as foragers came together from wide-rangingareas to directly acquire stone before dispersing again. Alterna-tively, long-distance acquisition of stone via individual or groupintermediaries clearly implies awidespread social network inwhichthe exchange of physical resources occurred between mobilehunteregatherer bands (Nash et al., 2013; Speth et al., 2013). Ter-ritories were “permeable” in that high-quality resources wereavailable to all, and groups likely did not restrict resources in onearea from groups originating from another area (Speth et al., 2013).

A significant obstacle to demonstrating Clovis long-distanceacquisition of stone, however, is being able to empirically andobjectively link the stone found at an archaeological site to aparticular outcrop (Burke, 2006; Burke et al., 2014; Haynes, 2002;Nash et al., 2013; Speer, 2014). Virtually all raw-material identifica-tions of Clovis tools continue to be based on qualitative descriptionsof megascopic and macroscopic properties. The problem with thisapproach is that such properties alone are insufficient for discrimi-nating among similar materials or for assigning specific provenanceto materials if they occur in geographically widespread outcrops(Hoard et al., 1992; Huckell et al., 2011; Luedtke, 1979, 1992; Speer,2014). Further, the identification of a stone raw material basedsolely on the appearance of a specimen in hand is subjective andoftentimes difficult to verify. As Luedtke (1993: 56e57) put it, “we alllearn to recognize lithic types through an unsystematic learningprocess that does not include any tests to verify the accuracy of ourlearning…[T]he relationship between our typologies and geologicalreality is completely untested and unknown.” When mega- andmacroscopic identifications are tested for accuracy, the results areoften surprisingly poor (Calogero, 1992; Ives, 1984, 1986; Luedtke,1992, 1993; Nance, 1984). The problem is not the use of qualitativetraitsdsystematic studies of qualitative traits such as fluorescenceunder ultraviolet light, color, texture, and presence of inclusions mayprovide reasonably confident provenance assignments (e.g., Hofmanet al., 1991; Luedtke 1992; Lyons et al., 2003; Malyk-Selivanova et al.,1998)dbut few such systematic studies have been undertaken.Rather, the problem arises from a conflation of extensionally deriveddescriptional ideational units, i.e., “types” of stone within a folkgeological taxonomy, with empirical units, i.e., stone available from aspecific locationda problem not unique to lithic-sourcing studies inarchaeology (O'Brien and Lyman, 2002).

One alternative is trace-element analysis, a method commonlyused on obsidian artifacts because of the chemical homogeneity ofobsidian and the nearly unique chemistry of individual obsidiansources (Amick, 1997; Boulanger et al., 2014; Cannon and Hughes,1997; Hester, 1988; Janetski and Nelson, 1999; Johnson et al., 1985;Kunselman, 1991; Smith and Kielhofer, 2012). However, in nearlyall these instances obsidian has been either minority components ofassemblages or isolated finds (e.g., Mockingbird Gap, New Mexico

(Hamilton et al., 2009, 2013) and Murray Springs, Arizona (Shackley,2007)), although one important exception to this pattern is the Dietzsite in Oregon, where nearly all the Clovis artifacts are made fromobsidian that has been geochemically sourced to outcrops no morethan 120 km away (Beck and Jones, 2013; Pinson, 2011).

Determining the source of tools made from chert and othercryptocrystalline silicatesdthe preferred raw materials of Clovisknappersdis more challenging because of the tendency of chert tobe heterogeneous within sources and, at times, to overlap inchemical composition among sources. Until the study reported here,the majority chert toolstone from only one unequivocal Clovis site inwestern North America had received chemical characterization: theEckles site, a Clovis residential occupation in north-central Kansas.There, stone tools were shown to be made predominately (75e80%)from White River Group silicate, the source of which is in northeastColorado, 450 km west of Eckles (Hoard et al., 1992, 1993; Holen,2010). Despite this lone empirical datum point, the assumption offrequent Clovis long-distance acquisition of stone is widely held (e.g.,Bradley et al., 2010; Gramly, 1999; Meltzer, 2009; Speth et al., 2013;Tankersley, 2004). Herewe assess the assumption using geochemicaldata acquired through neutron activation analysis (NAA) of stoneflaking debris from a Clovis site in eastern North America, PaleoCrossing (33ME274), located in Medina County, Ohio. (Fig. 1).

Paleo Crossing provides a unique opportunity to assess long-distance acquisition because the majority of the site's Clovis arti-facts (>50%) have been proposed to bemade fromWyandotte chert,based solely on visual examination of the entire assemblage (Brose,1994; Tankersley and Holland, 1994). The closest sources ofWyandotte chert are in Harrison and Crawford counties, Indianadastraight-line distance from Paleo Crossing of some 450e510 km(Brose, 1994; Eren, 2006, 2010; Eren and Redmond, 2011; Erenet al., 2004, 2005; Miller, 2013, 2014; Tankersley and Holland,1994). It is important to note that we are not questioning previ-ous macroscopic analyses suggesting that there is a majority stoneraw material at Paleo Crossing visually consistent with Wyandottechert. By employing NAAwe are testing whether that majority rawmaterial is in fact Wyandotte chert, rather than a “look alike.”

This 450e510 km is an extreme source-to-site distance pro-posed for a Clovis “exotic” assemblage in eastern North America(Ellis, 2011). In fact, it is a minimum estimate of acquisition distancebecause foragers do not travel in straight lines (Meltzer, 2009). Aleast-cost path based on slope indicates the distance betweenWyandotte chert sources and Paleo Crossing is 825 km (Fig. 1, routeA), whereas different river routes are over 1000 km (Fig. 1, routes B& C). The assemblage from Paleo Crossing also includes materialvisually consistent with sources located less than 200 km from thesite, such as Flint Ridge and Upper Mercer (Fig. 1) (Brose, 1994;Tankersley and Holland, 1994).

Between 1990 and 1993, staff from the Department of Archae-ology at the Cleveland Museum of Natural History (CMNH), alongwith associated researchers, carried out survey and test excavationsat the site. The project was stimulated by the 1989 discovery ofClovis projectile points and other stone tools by a local collector,who contacted the CMNH. Radiocarbon dating of charcoal from asubsurface feature provided a date of 10,980 ± 75 14C YBP1 (Brose,

Fig. 1. Location of the Paleo Crossing Clovis site relative to the broader North American Midwest and Great Lakes regions, other Ohio Clovis sites, and the Wyandotte chert sourcearea in south-central Indiana (1). Note that there are several toolstone sources closer to Paleo Crossing than the Wyandotte source, including Vanport and Brush Creek (2), BrushCreek (3), Upper Mercer (4), Flint Ridge (5), Plum Run (6), Pipe Creek (7), and Ten Mile Creek (8). Although the straight-line distance between Paleo Crossing and the Wyandottechert source is 450e510 km, this linear distance should be considered a minimum estimate. A least-cost path derived from slope is 825 km (A), while an Ohio RivereMuskingumRivereKillibuck River route is 1275 km (B) and an Ohio RivereMuskingum RivereTuscarawas River route is 1240 km (C).

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558552

1994) (calibrated at 2-sigma to 13,023e12,717 calBP using OxCal4.2 (Bronk Ramsey, 2009) and IntCal13 (Reimer et al., 2013)). Theearliest Paleoindian fluted projectile points in the Northeast havebeen often argued to be a post-Clovis cultural manifestation called“Gainey,” although geometric morphometric analyses of pointsfrom Paleo Crossing (Fig. 2aed) and “classic” Clovis points from theSouthern Plains and Southwest show that Paleo Crossing points areindistinguishable from points typed as Clovis (Buchanan et al.,2014; see also Smith et al., in press) (see Supplementary OnlineMaterials). The site has yielded other characteristic Clovis arti-facts, including endscrapers, some with spurs and notches(Fig. 2eeg); graver spurs (Fig. 2h); evidence for prismatic bladeproduction (Fig. 2i); overshot flakes2 (Fig. 2j); and overshot mis-takes on early stage fluted bifacial preforms (Fig. 2k).

2 Although recent archaeological and experimental analyses have suggestedovershot flakes to be mistakes, i.e., there was not an “intentional” or “controlledovershot flaking strategy” (Eren et al., 2013), it is possible that the presence ofovershot flakes, or an increased frequency of overshot flakes, is still diagnostic ofClovis (Eren et al., 2014). That is why we believe it is important to mention thepresence of overshot flakes at Paleo Crossing. However, the reader should note thatMu~niz (2014) recently showed that there were no significant differences in over-shot flaking between bifaces from Clovis caches and bifaces from Late Prehistoriccaches. If further analyses support this result, then the presence of overshot flakesor overshot flake scars, even as mistakes, may not be diagnostic of Clovis.

2. Materials and methods

2.1. Sample

Eighty-eight stone artifacts from Paleo Crossing visuallyconsistent with Wyandotte chert were selected and sent to theUniversity of Missouri Research Reactor (MURR). These artifactsconsisted of small flakes, broken flakes, and block shatter resultingfrom stone-tool production and tool resharpening, with an averagemass of 1.7 g and standard deviation of 2.38 g. Given that NAA is adestructive analysis, we chose chert specimens from Paleo Crossingthat were collected by the landowner and were without specificprovenience beyond the site itself. Thus, beyond visual criteria, forall intents and purposes the sample was essentially selected atrandom. Of these, 34 were compositionally analyzed by NAA. Theremaining fifty-four archaeological specimens were not necessaryfor NAA analysis and were saved for future archaeological analysis.An additional 30modern specimens ofWyandotte chert taken fromnodules from Harrison County, Indiana, were also sent to MURR(Figs. 3e4).

2.2. Sample preparation

All specimens were washed in deionized water and scrubbedwith a stiff-bristle brush to remove detritus from their surfaces.

Fig. 2. Characteristic Clovis artifacts from Paleo Crossing, Ohio, include fluted projectile points (a, b, c, d), endscrapers (e, f, g), graver spurs (h), a prismatic blade core remnant (i),overshot flakes (j), and overshot mistakes on fluted bifacial preforms (k). All artifacts are visually consistent with Wyandotte chert, except for (h top), which is visually consistentwith Upper Mercer chert, and (k), which is visually consistent with Flint Ridge chalcedony.

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558 553

Inked catalog numbers were removed from artifacts with ethanol.Small artifact specimenswere broken in an agatemortar and pestle,whereas the larger source specimens were placed between twotool-steel plates and crushedwith a Carver Press. Small, 50e100mgfragments were obtained from each crushed specimen, and thesewere examined under low-power magnification. Fragments withmetallic streaks or crush fractures were removed from the sample.

Fig. 3. Wyandotte chert nodules (often referred to as “cannonballs”) from HarrisonCounty, Indiana.

Two analytical samples were prepared from each specimen. Por-tions of approximately 200 mg of rock fragments were weighedinto high-density polyethylene vials used for short irradiations atMURR. At the same time, 800mg aliquots from each specimenwereweighed into high-purity quartz vials used for long irradiations.Individual sample weights were recorded to the nearest 0.01 mgusing an analytical balance. Along with the unknown samples,standards made from National Institute of Standards and Tech-nologyecertified standard reference materials of SRM-1633b (CoalFly Ash), SRM-278 (Obsidian Rock), and SRM-688 (Basalt Rock)were similarly prepared.

2.3. Irradiation and gamma-ray spectroscopy

NAA at MURR consists of two irradiations and three gammacounts (Glascock, 1992; Glascock and Neff, 2003; Neff, 2000). Ashort (5 s) irradiation is performed through a pneumatic-tubesystem. Samples in the polyvials are sequentially irradiated, twoat a time, for 5 s by a neutron flux of 8 � 1013 n cm�2 s�1. The 720-s count yields gamma spectra containing peaks for Al, Ba, Ca, Dy, K,Mn, Na, Ti, and V. The long-irradiation samples in high-purityquartz vials and are subjected to a 70-h irradiation at a neutronflux of 5 � 1013 n cm�2 s�1. Samples decay for seven days afterirradiation and are then counted for 1800 s on a high-resolution Gedetector coupled to an automatic sample changer. This count pro-vides determinations of seven elements: As, La, Lu, Nd, Sm, U, and

Fig. 4. Wyandotte chert nodule from Harrison County, Indiana, showing a fresh frac-ture and the tool-stone beneath the outer cortex.

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558554

Yb. After an additional three- or four-week decay, a final count of8500 s is performed on each sample, which provides yield de-terminations for 17 elements: Ce, Co, Cr, Cs, Eu, Fe, Hf, Ni, Rb, Sb, Sc,Sr, Ta, Tb, Th, Zn, and Zr.

2.4. Data analysis and comparative data

Comparative data used in this study are drawn from previousresearch at MURRdWyandotte chert (Harrison Co., IN) and Plum-mer (aka Lead Creek) chert (Glascock, 2004) and CobdeneDongolachert from Union Co., IL (Morrow et al., 1992)dand from unpub-lished data in the MURR database for Flint Ridge (Licking Co., OH),Upper Mercer (Coshocton and Perry Cos., OH), and Brush Creek/Hughes River (Ritchie Co., WV). Statistical analyses were performedon base-10 logarithms of calculated elemental abundances. Statis-tical analyses followed standard multivariate methods employed inthe use of compositional data to assess provenance (Baxter andBuck, 2000; Bieber et al., 1976; Bishop and Neff, 1989; Harbottle,1976; Neff, 2002).

3. Results

Nineteen of the 34 (56%) Paleo Crossing debitage specimensexhibit elevated levels of U and Sb, which distinguishes them fromall other blue-gray chert specimens in the MURR database exceptfor those obtained from the Ste. Genevieve Formation (seeSupplementary Online Materials). Based on concentrations of thesetwo elements alone, all other midwestern sources of blue-graychert may be eliminated as potential sources of the 34 pieces ofPaleo Crossing debitage. Of the two major sources of chert withinthe Ste. Genevieve Formation, the Paleo Crossing artifacts exhibitconcentrations of Al, Ca, Ti, and V that place them only within therange documented for Wyandotte chert from Harrison Co., Indiana(Fig. 5). We note that the Paleo Crossing artifacts plot on the edgesof the 90% confidence ellipse for our Wyandotte source samplewhen projected in canonical discriminant space (e.g., Fig. 5). Thissuggests to us that the specimens comprising our Wyandottesource sample likely do not fully represent the chemical variabilitypresent within chert from the Ste. Genevieve formation. Thisshould come as no surprise given that the Ste. Genevieve Formationis mapped throughout much of western Indiana from the OhioRiver northward to Putnam Co., yet all of our source specimensderive from exclusively from Harrison Co.

In most projections of the data, the remaining 15 (44%) speci-mens show affinity with Vanport Formation chert from the Flint

Ridge quarries of Ohio. These archaeological specimens, however,have consistently low to moderate probabilities of belonging to acompositional group matching MURR's Flint Ridge comparativespecimens, which exhibit marked compositional heterogeneity.One possibility is that the 15 archaeological specimens derive fromFlint Ridge but MURR's current non-artifactual comparative sampledoes not fully represent the range of chemical variation of thesource. The other possibility is that the specimens derive from asource of Vanport Formation chert not represented in the MURRdatabase. In either case, the remaining 15 archaeological specimensderive from the same non-Wyandotte geochemical source that ismost chemically similar to Flint Ridge (Fig. 5).

4. Discussion

The results of the first reported geochemical analyses of Clovisartifacts from eastern North America support the hypothesis thatthe majority of Paleo Crossing artifacts are made from Wyandottechert (Brose, 1994; Tankersley and Holland, 1994), the primarysource of which is 450e510 linear km away from the site. Given thatthe majority chert toolstone from only one Clovis site in westernNorth AmericadEckles, Kansas (Hoard et al., 1992, 1993; Holen2010)dhas been geochemically verified to demonstrate long-distance stone acquisition, our results from eastern North Americamore broadly support the long-held but poorly corroboratedassumption that Clovis foragers regularly depended on “exotic”stone for tool production. Furthermore, long-distance stoneacquisition, either through direct or indirect acquisition of thestone, is alsowidely acknowledged to be evidence of geographicallywidespread social contact and territorial permeability (Bamforth,2009; Ellis, 2011; Goodyear, 1989; Pearce, 2014; Pearce andMoutsiou, 2014; Sholts et al., 2012; Speth et al., 2013). PaleoCrossing thus provides strong inferential material evidence that thefast expansion of the Clovis culture, as already documented bychronometric evidence (Prasciunas and Surovell, in press; Watersand Stafford, 2007), occurred as a result of a geographically wide-spread hunteregatherer social network.

Other asserted claims for long-distance Clovis stone acquisitionin eastern North America based on macroscopic evaluation of rawmaterial, while not validated, can now at least be consideredplausible. Our results thus in part substantiate the recent work ofEllis (2011), who compared Clovis stone-acquisition distance to thatof post-Clovis foragers. His meta-analysis, based on visual inspec-tion of materials from 83 Paleoindian sites (including PaleoCrossing) in the Great Lakes, New England, Middle Atlantic, andCanadian Maritimes regions, showed Clovis to possess significantlyhigher stone-acquisition distances than that of post-Clovis foragers.Given that Clovis represents the earliest significant archaeologicalsignal in northeastern North America, Ellis's (2011) work, especiallyin light of our results, is consistent with the predictions of foragercolonization theory, specifically that Clovis foragers “more closelytied to the colonization process” (2011, p. 387) would have main-tained contact among dispersed groups in order to sustain infor-mation flow, social relations, and demographic viability on alandscape in which their populations were low and otherwisethinly scattered (Ellis, 2008; Sholts et al., 2012; Brantingham, 2006;Meltzer, 2004; Surovell, 2000). Nevertheless, Ellis's conclusionswould be considerably strengthened through geochemical analysistesting of the assemblages he included.

Based on these results we propose that direct human acquisitionand group transport of Wyandotte chert is the most parsimoniousexplanation for its presence at Paleo Crossing (see Ellis, 2011;Meltzer, 1989), although we cannot rule out indirect acquisitionthrough intermediaries or groups or direct acquisition throughlogistical forays by individuals (e.g., Speth et al., 2013). Given that

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558 555

there are approximately 500 Clovis tools and over 10,000 pieces ofdebitage from Paleo Crossing (Barrish, 1995; Eren, 2011), and onlyan estimated 5% of the site was excavated (Brose, 1994), Wyandottechert is not only the majority raw material in the assemblage butalso represents a substantial amount of stone in terms of weight.3 Itseems unlikely that Clovis foragers residing at Paleo Crossing, andwho found themselves without their majority raw material, wouldhave acquired it indirectly from a source 450e510 km away whileignoring several sources located less than 100 km away (e.g., Burke,2006; Ellis, 2011). The same rationale can be applied to the acqui-sition of Wyandotte chert by means of a logistical foray by aspecialized task group. Instead, a more parsimonious explanation isthat Clovis foragers “geared up” with Wyandotte chert beforeentering and exploring the southern Great Lakes region, which wasdeglaciated after 14,000 B.P. (ca. 17,000 calBP), with ecologicallyproductive environments in place perhaps by 13,000e12,500 B.P.(ca. 15,000 calBP) (Gill et al., 2012; Glover et al., 2011; Herdendorf,2013; Hill, 2006; Metcalfe et al., 2013; Szabo and Chanda, 2004; Yu,2000). After which these foragers could have procured stone fromoutcrops located closer to Paleo Crossing either during theirjourney to the site or through exchange or logistical forays afterestablishing a camp at Paleo Crossing. This, of course, is not to saythat Clovis colonizing foragers had no knowledge of the areawhatsoever. Instead, it is entirely possible that a “working knowl-edge” of the area had been hitherto achieved through earlierscouting trips aimed to target desirable resource areas (Dincauze,1993; Ellis, 2011; Eren, 2011; Eren et al., 2012).

If we take the conjecture of direct acquisition (gearing up) a bitfurther, we note that Sheriden Cave is not far from the least-costpath between the Wyandotte chert source area and Paleo

3 Researchers have suggested that including both percentages and weights ofstone raw material should be standard practice in Paleoindian studies (Bamforth,2009; Speth et al., 2013). With respect to weight, Speth et al. (2013:121) state,“There clearly is a great need for more data on the actual weight of flint, both localand non-local, that is recovered in Paleoindian sites. Without such data, argumentsabout the mechanisms by which raw materials moved from quarry to site willremain severely handicapped.”

With respect to Paleo Crossing, Brose (1994:66) stated that “initial appraisal ofthe chipped stone tools' lithology displayed a remarkable predominance of exoticchert” and suggested that “over 65% of the tools are chipped of Wyandotte chert.”Tankersley and Holland (1994:61e62) suggested that 158 (79%) of “more than 200distinctive early Paleoindian chipped-stone artifacts … are manufactured from aMississippian-age, upper Ste. Genevieve formation chert [and that] Wyandotte isthe closest source of this material.” Although not stated explicitly, the affinity withWyandotte chert was likely made by comparisons of color, texture, and fossil in-clusions with source specimens under low- and medium-power magnification (e.g.,Tankersley, 1984, 1989). Eren's analysis of the entire Paleo Crossing assemblagesuggested that 67.6% (n ¼ 23) of the fluted points (Eren et al., 2004), 73% (n ¼ 151)of unifacially worked scrapers (Eren et al., 2005), and 75.8% (n ¼ 22) of non-projectile point bifaces (Eren, 2006) are made on stone that is visually consistentwith hand samples of Wyandotte chert curated at the Cleveland Museum of NaturalHistory.

How do these percentages translate into weights? Analyses of the PaleoCrossing stone-tool assemblage are still ongoing, but we can make some estimates.Following Bamforth (2009:142), we focus here on the unifacial stone tools, giventhat projectile points differ from most other contemporaneous tools in ways thatmake unreliable sources of information on mobility. The 401 unifacial stone tools atPaleo Crossing collectively weigh 2.54 kg (5.6 lbs). Seventy-three percent of theseare made from Wyandotte chert, 1.85 kg (4.08 lbs). This may not seem like a lot ofweight, but only 5% or so of Paleo Crossing was excavated (Barrish, 1995). Thus,1.85 kg/0.05 ¼ 37 kg (81 lbs). This estimation, however, does not consider that (1)most (n ¼ 239) of the unifaces included in this calculation are broken, and there areno known refits. Thus, complete uniface specimens would substantially increasethis calculation; and (2) the calculation does not include any weights from flutedpoints, bifaces/preforms, or debitage, of which there are over 10,000 specimens ofthe latter. Thus, although our simple extrapolation assumes a uniform distributionof Wyandotte unifacial tools across the site as well as a consistent proportion ofthese items across the site, 37 kg (81 lbs) of Wyandotte toolstone at Paleo Crossingmay still very well be a minimum estimate given the two stipulations above.

Crossing (Fig. 1). Sheriden Cave, which possessed two bone pro-jectile points, a Clovis fluted point, and a thinned flake-blank largeenough to be turned into a fluted point, could conceivably beinterpreted as a cache (Redmond and Tankersley, 2005). We find itinteresting that not only are the dates of Sheriden Cave (10,915 ± 3014C YBP (Waters et al., 2009), calibrated at 2-sigma range to12,821e12,708 calBP), virtually identical to that of Paleo Crossing(10,980 ± 75 14C YBP, calibrated at 2-sigma to 13,023e12,717 calBP),the flake blank is made on stone visually consistent with, andasserted to be, Wyandotte chert (Redmond and Tankersley, 2005).We are not suggesting that Sheriden Cave and Paleo Crossing arenecessarily connected, but the preponderance of Clovis cachesacross North America has been linked to the landscape unfamil-iarity inherent to the colonization process, meaning that Clovisgroups new to the landscape did not know where or when theywere next going to find a suitable stone outcrop. By depositingcaches or insurance gear, “they created artificial resupply depots,and thus anticipated and compensated for that lack of knowledge”(Meltzer, 2009, p. 252). Thus, if indeed caching, or increased fre-quency of caching, can be linked to colonization behavior, thepresence of caching in eastern North America is consistent with ourscenario that the Clovis group or groups that occupied PaleoCrossing geared up with Wyandotte chert prior to heading north-east toward the Great Lakes, into a regionwhere stone sources wereinitially unknown.

NAA results also tentatively validate the provenance of one ofthe hypothesized secondary raw materials at Paleo Crossing, FlintRidge chalcedony (Tankersley and Holland, 1994), located approx-imately 140 km south of the site. Given that all of the lithic artifactssubmitted for geochemical analyses were visually consistent withWyandotte chert, this unexpected result strongly reaffirms theneed for more regular use of geochemical analyses to determinestone provenance and inferences of Clovis long-distance toolstoneacquisition. However, this is not to say that Wyandotte chert andFlint Ridge chalcedony are macroscopically similar when exam-ining larger lithic specimens like large flakes, fluted points, andendscrapers. Indeed, on sizable specimens these two raw materialsare easy to distinguish visually. However, visual identification be-comes commensurately harder as lithic specimens become smaller.Given that our tested artifacts were on average only 1.7 g, perhaps itshould come as no surprise that some of the chips submitted fortesting were consistent with a non-Wyandotte stone. In otherwords, if we had submitted tools rather than small flaking debrisfor NAA testing, our sample percentage may very well have been, orclose to, 100% consistent with Wyandotte chert. To reiterate ourstatement of purpose from the introduction, we were not ques-tioning previous macroscopic analyses suggesting that there is amajority stone raw material at Paleo Crossing visually consistentwith Wyandotte chert. Instead, by employing NAA we were testingwhether that majority raw material is in fact Wyandotte chertrather than a “look alike.”

If Flint Ridge chalcedony was acquired directly by foragersduring the journey between the Wyandotte source area and PaleoCrossing, then this scenario would be consistent with the model offoragers “topping-up” their stone supply rather than replacing theirsupply entirely with a newly encountered stone (Brantingham,2006). Nevertheless, it is also possible that Flint Ridge stone wasprocured during a logistical foray after Clovis foragers arrived atPaleo Crossing (Morgan et al., in press) or that the stone arrived atPaleo Crossing through intermediaries (e.g., Speth et al., 2013).Given that only 5% of the visually analyzed Clovis tools wereassigned to Flint Ridge chalcedony (Brose, 1994; Tankersley andHolland, 1994), it is surprising that there seems to be dispropor-tionately more flaking debris. Perhaps Clovis foragers “geared up”with tools made of Flint Ridge chalcedony and thus removed them

Fig. 5. Plot of the first two canonical discriminant functions showing blue-gray chert source samples and artifacts from Paleo Crossing. Ellipses are drawn at the 90% confidenceinterval.

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558556

from the site, leaving the debris behind (Morgan et al., in press). Wefind it interesting that the closest Clovis site to Paleo Crossing, andone of the largest Clovis sites in North America, Nobles Pond (Fig.1),not only exhibits 40e45% stone visually consistent with Flint Ridgechalcedony (Seeman, 1994) but also shows a small amount of stonevisually consistent with Wyandotte chert (Seeman et al., 2013).Similar to our discussion of Sheriden Cave, we are not suggestingthat Nobles Pond and Paleo Crossing are necessarily connected sodirectly, only that, if shown to be true, two sites each exhibitinginverse primary and secondary raw materials is consistent with ascenario of stone exchange.

It is important to note, however, that there is evidence for asmall Archaic occupation at Paleo Crossing (Brose, 1994; Eren andKollecker, 2004). While only Clovis artifacts are manufactured onstone visually consistent with Wyandotte chert, both Clovis andArchaic projectile points are manufactured on stone visuallyconsistent with Flint Ridge chalcedony (Brose, 1994; Eren andKollecker, 2004). Given that small flakes, broken flakes, and blockshatterdthe artifact types tested here with NAAdare not currentlytemporally diagnostic, it is impossible for us to know at the presenttime whether the Flint Ridge flaking debris at Paleo Crossing ispredominately Clovis or Archaic or instead a mixture of both.However, the fact that Archaic projectile points are made of FlintRidge may go a long way toward explaining the discrepancy be-tween the smaller percentage of Clovis tools and the larger per-centage of flaking debris made on that stone type.

Acknowledgments

We thank David Meltzer for providing constructive commentson an early draft of this paper, improving it substantially. We alsothank Richard Klein, Christopher Ellis, and four anonymous re-viewers for their constructive comments. Analyses were performedat the Archaeometry Laboratory at MURR, which is supported by agrant from the National Science Foundation (BCS #1110793). M.I.E.

is financially supported by a University of Missouri College of Artsand Science postdoctoral fellowship, and is grateful for the supportof Rebecca Catto and Mustafa, Kathleen, and Nimet Eren.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jas.2014.11.005.

References

Amick, D., 1997. Geochemical source analysis of obsidian Paleoindian points fromthe Black Rock Desert, Nevada. Curr. Res. Pleistocene 14, 97e99.

Bamforth, D., 2009. Projectile points, people and plains Paleoindian perambula-tions. J. Anthropol. Archaeol. 28, 142e157.

Barrish, B., 1995. The Paleo Crossing Site: Fluted Point Typology and Chronology(M.A. thesis). University of Kent, Kent, Ohio.

Baxter, M., Buck, C., 2000. Data handling and statistical analysis. In: Ciliberto, E.,Spoto, G. (Eds.), Modern Analytical Methods in Art and Archaeology, ChemicalAnalysis Series, vol. 151. Wiley, New York, pp. 681e746.

Beck, C., Jones, T., 2013. In: Graf, K., Ketron, C., Waters, M. (Eds.), Complexities of theColonization Process: a View from the North American West. PaleoamericanOdessey, Texas A&M University, pp. 273e291.

Bieber Jr., A., Brooks, D., Harbottle, G., Sayre, E., 1976. Application of multivariatetechniques to analytical data on Aegean ceramics. Archaeometry 18, 59e74.

Bishop, R., Neff, H., 1989. Compositional data analysis in archaeology. In: Allen, R.(Ed.), Archaeological Chemistry IV, Advances in Chemistry Series No. 220.American Chemical Society, Washington, DC, pp. 57e86.

Boulanger, M., Glascock, M., Shackley, M., Skinner, C., Thatcher, J., 2014. Likelysource attribution for a possible Paleoindian obsidian tool from northwestLouisiana. La. Archaeol. 37, 87e94.

Bradley, B., Collins, H., Hemmings, A., 2010. Clovis Technology. InternationalMonographs in Prehistory, Ann Arbor.

Brantingham, P., 2006. Measuring forager mobility. Curr. Anthropol. 47, 435e459.Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51 (1),

337e360.Brose, D., 1994. Archaeological investigations at the Paleo Crossing site, a Paleo-

indian occupation in Medina County, Ohio. In: Dancey, W. (Ed.), The First Dis-covery of America: Archaeological Evidence of the Early Inhabitants of the OhioArea. Ohio Archaeological Council, Columbus, pp. 61e76.

Buchanan, B., O'Brien, M., Collard, M., 2014. Continent-wide or region-specific? Ageometric morphometrics-based assessment of variation in Clovis point shape.Archaeol. Anthropol. Sci. 6, 145e162.

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558 557

Burke, A., 2006. Paleoindian ranges in northeastern North America based on lithicraw materials sourcing. In: Bressy, C., Burke, A., Chalard, P., Martin, H. (Eds.),Notions de territoire et de mobilit�e. Exemples de l’Europe et des premi�eresnations en Am�erique du Nord avant le contact europ�een, Actes de sessionspr�esent�ees au Xe congr�es annuel de l’Association Europ�eenne des Arch�eologues(EAA), Lyon, 8e11 Septembre 2004, pp. 77e89.

Burke, A., Gauthier, G., Chapdelaine, C., 2014. Refining the Paleoindian lithic sourcenetwork at Clich�eeRancourt using XRF. Archaeol. East. N. Am. 41, 101e128.

Calogero, B.L., 1992. Lithic misidentification. Man Northeast 43, 87e90.Cannon, K., Hughes, R.E., 1997. Provenance analysis of obsidian Paleoindian pro-

jectile points from Yellowstone National Park, Wyoming. Curr. Res. Pleistocene14, 101e104.

Dincauze, D.F., 1993. Pioneering in the Pleistocene: large Paleoindian sites in theNortheast. Archaeol. East. N. Am. 25, 43e60.

Ellis, C., 1989. The explanation of Northeastern Paleoindian lithic procurementpatterns. In: Ellis, C., Lothrop, J. (Eds.), Eastern Paleoindian Lithic Resource Use.Westview, Boulder, pp. 139e164.

Ellis, C., 2008. The Fluted Point tradition and the Arctic Small Tool tradition: what'sthe connection? J. Anthropol. Archaeol. 27, 298e314.

Ellis, C., 2011. Measuring Paleoindian range mobility and land-use in the GreatLakes/Northeast. J. Anthropol. Archaeol. 30, 385e401.

Eren, M.I., 2006. The Paleo Crossing (33-ME-274) non-projectile point bifaceassemblage. Curr. Res. Pleistocene 23, 95e97.

Eren, M.I., 2010. Anvil reduction at the Early Paleoindian site of Paleo Crossing(33ME274), Northeast Ohio. Curr. Res. Pleistocene 27, 75e77.

Eren, M.I., 2011. Behavioral Adaptations of Late Pleistocene Human Colonizers in theNorth American Lower Great Lakes Region (Ph.D. dissertation). SouthernMethodist University, Dallas, TX.

Eren, M.I., Kollecker, M.A., 2004. Data on the post-Paleoindian projectile pointassemblage from the Paleo Crossing Site (33-ME-274). Ohio Archaeol. 54, 10e11.

Eren, M.I., Redmond, B., 2011. Clovis blades at Paleo Crossing (33ME274), MedinaCounty, Ohio. Midcont. J. Archaeol. 36, 173e194.

Eren, M.I., Redmond, B., Kollecker, M., 2004. The Paleo Crossing (33-ME-274) flutedpoint assemblage. Curr. Res. Pleistocene 21, 38e39.

Eren, M.I., Redmond, B., Kollecker, M., 2005. Unifacial stone tool analyses from thePaleo Crossing Site (33-ME-274), Ohio. Curr. Res. Pleistocene 22, 43e45.

Eren, M.I., Chao, A., Hwang, W., Colwell, R., 2012. Estimating the richness of apopulation when the maximum number of classes is fixed: a nonparametricsolution to an archaeological problem. PLOS ONE 7 (5), e34179.

Eren, M.I., Patten, R.J., O'Brien, M.J., Meltzer, D.J., 2013. Refuting the technologicalcornerstone of the Ice-Age Atlantic crossing hypothesis. J. Archaeol. Sci. 40,2934e2941.

Eren, M.I., Patten, R.J., O'Brien, M.J., Meltzer, D.J., 2014. More on the rumor of“intentional overshot flaking” and the purported Ice-Age Atlantic crossing.Lithic Technol. 39, 55e65.

Ferring, C., 2001. The Archaeology and Paleoecology of the Aubrey Clovis Site(41DN479). University of North Texas, Denton Center for EnvironmentalArchaeology, Denton County, Texas.

Frison, G., Bradley, B., 1999. The Fenn Cache: Clovis Weapons and Tools. One HorseLand and Cattle Company, Santa Fe, NM.

Gill, J.L., Williams, J.W., Jackson, S.T., Donnelly, J.P., Schellinger, G.C., 2012. Climaticand megaherbivory controls on late-glacial vegetation dynamics: a new, high-resolution, multi-proxy record from Silver Lake, Ohio. Quat. Sci. Rev. 34,66e80.

Gingerich, J., 2011. Down to seeds and stones: a new look at the subsistence remainsfrom Shawnee-Minisink. Am. Antiq. 76, 127e144.

Glascock, M., 1992. Characterization of archaeological ceramics at MURR by neutronactivation analysis and multivariate statistics. In: Neff, H. (Ed.), ChemicalCharacterization of Ceramic Pastes in Archaeology, vol. 7. Prehistory Press,Madison, pp. 11e26.

Glascock, M., 2004. Neutron activation analysis of chert artifacts from a Hopewellmound. J. Radioanal. Nucl. Chem. 262, 97e102.

Glascock, M., Neff, H., 2003. Neutron activation analysis and provenance research inarchaeology. Meas. Sci. Technol. 14, 1516e1526.

Glover, K.C., Lowell, T.V., Wiles, G.C., Pair, D., Applegate, P., Hajdas, I., 2011. Degla-ciation, basin formation and post-glacial climate change from a regionalnetwork of sediment core sites in Ohio and eastern Indiana. Quat. Res. 76,401e410.

Goebel, T., Waters, M., O'Rourke, D., 2008. The late Pleistocene dispersal of modernhumans in the Americas. Science 319, 1497e1502.

Goodyear, A., 1989. A hypothesis for the use of cryptocrystalline raw materialsamong Paleoindian groups of North America. In: Ellis, C., Lothrop, J. (Eds.),Eastern Paleoindian Lithic Resource Use. Westview Press, Boulder, pp. 1e9.

Gramly, R., 1999. The Lamb Site: a Pioneering Clovis Encampment. PersimmonPress, North Andover, MA.

Hamilton, M.J., Buchanan, B., 2007. Spatial gradients in Clovis-age radiocarbon datesacross North America suggest rapid colonization from the north. Proc. Natl.Acad. Sci. 104, 15625e15630.

Hamilton, M.J., Huckell, B., Shackley, M.S., 2009. Clovis Obsidian sources in theCentral Rio Grande Rift region of New Mexico. Curr. Res. Pleistocene 26, 62e65.

Hamilton, M.J., Buchanan, B., Huckell, B., Holliday, V.T., Shackley, M.S., Hill, M.E.,2013. Clovis Paleoecology and lithic technology in the Central Rio Grande RiftRegion, New Mexico. Am. Antiq. 78, 248e265.

Harbottle, G., 1976. Activation analysis in archaeology. In: Newton, G.W.A. (Ed.),Radiochemistry, vol. 3. The Chemical Society, London, pp. 33e72.

Hayden, B., 1982. Interaction parameters and the demise of Paleo-Indian crafts-manship. Plains Anthropol. 27, 109e123.

Haynes, C.V., Donahue, D., Jull, A., Zabel, T., 1984. Application of accelerator dating tofluted point Paleoindian sites. Archaeol. East. N. Am. 12, 184e191.

Haynes, G., 2002. The Early Settlement of North America: the Clovis Era. CambridgeUniversity Press, Cambridge.

Haynes, G., Anderson, D., Ferring, C., Fiedel, S., Grayson, D., Haynes, C.V., Holliday, V.,Huckell, B., Kornfeld, M., Meltzer, D., Morrow, J., Surovell, T., Waguespack, N.,Wigand, P., Yohe, R., 2007. Comment on “Redefining the Age of Clovis: impli-cations for the Peopling of the Americas”. Science 317, 320b.

Herdendorf, C.E., 2013. Research overview: Holocene development of Lake Erie.Ohio J. Sci. 112, 24e36.

Hester, T.R., 1988. Paleoindian obsidian artifacts from Texas: a review. Curr. Res.Pleistocene 5, 27e29.

Hill, C.L., 2006. Geologic framework and glaciation of the eastern area. In:Ubelaker, D.H. (Ed.), Handbook of North American Indians, Environment, Ori-gins, and Population, vol. 3. Smithsonian Institution, Washington, pp. 81e98.

Hoard, R.J., Holen, S.R., Glascock, M.D., Neff, H., Elam, J.M., 1992. Neutron activationanalysis of stone from the Chadron Formation and a Clovis site on the GreatPlains. J. Archaeol. Sci. 19, 655e665.

Hoard, R.J., Bozell, J., Holen, S., Glascock, M., Neff, H., Elam, J., 1993. Source deter-mination of White River group silicates from two archaeological sites in theGreat Plains. Am. Antiq. 58, 698e710.

Hofman, J.L., Todd, L.C., Collins, M.B., 1991. Identification of central Texas Edwardschert at the Folsom and Lindenmeier sites. Plains Anthropol. 36, 297e308.

Holen, S., 2010. The Eckles Clovis site, 14JW4: a Clovis site in Northern Kansas.Plains Anthropol. 55, 299e310.

Holliday, V.T., 2000. The evolution of Paleoindian geochronology and typology onthe Great Plains. Geoarchaeology 15, 227e290.

Huckell, B., Kilby, J.D., Boulanger, M.T., Glascock, M.D., 2011. Sentinel Butte:neutron activation analysis of White River Group chert from a primary sourceand artifacts from a Clovis cache in North Dakota, USA. J. Archaeol. Sci. 38,965e976.

Ives, D.J., 1984. The Crescent Hills prehistoric chert quarrying area: more than justrocks. In: Butler, B., May, E. (Eds.), Prehistoric Chert Exploitation: Studies fromthe Midcontinent, Center for Archaeological Investigations. Southern IllinoisUniversity, Carbondale, pp. 187e196.

Ives, D.J., 1986. Hopewell interaction sphere “exotic” cherts: quantitative versusqualitative analyses. Trans. Am. Nucl. Soc. 53, 198e199.

Janetski, J.C., Nelson, F.C., 1999. Obsidian sourcing of Paleoindian points from theSevier Desert, Utah. Curr. Res. Pleistocene 16, 96e97.

Johnson, E., Holliday, V.T., Asaro, F., Stross, F.H., Michel, H.V., 1985. Trace elementanalysis of Paleoindian obsidian artifacts from the southern High Plains. Curr.Res. Pleistocene 2, 51e53.

Kelly, R., Todd, L., 1988. Coming into the country: early Paleoindian hunting andmobility. Am. Antiq. 53, 231e244.

Kilby, J.D., 2008. An Investigation of Clovis Caches: Content, Function, and Tech-nological Organization (Ph.D. dissertation). The University of New Mexico,Albuquerque.

Kunselman, R., 1991. Durable Paleoindian artifacts: XRF variability of selectedWyoming source materials. Wyo. Archaeol. 34 (1e2), 15e25.

Levine, M., 1990. Accommodating age: radiocarbon results and fluted point sites innortheastern North America. Archaeol. East. N. Am. 18, 33e63.

Luedtke, B.E., 1979. The identification of sources of chert artifacts. Am. Antiq. 44,744e756.

Luedtke, B.E., 1992. Archaeologist's Guide to Chert and Flint. Institute of Archae-ology, University of California, Los Angeles.

Luedtke, B.E., 1993. Lithic source analysis in New England. Bull. Mass. Archaeol. Soc.54, 56e60.

Lyons, W.H., Glascock, M.D., Mehringer, P.J., 2003. Silica from sources to site: ul-traviolet fluorescence and trace elements identify cherts from Lost Dune,southeastern Oregon, USA. J. Archaeol. Sci. 30, 1139e1159.

MacDonald, G., 1968. Debert: a Paleo-indian Site in Central Nova Scotia. PersimmonPress, Buffalo, NY.

Malyk-Selivanova, N., Ashley, G.M., Gal, R., Glascock, M.D., Neff, H., 1998. Geological-geochemical approach to sourcing of prehistoric chert artifacts, northwesternAlaska. Geoarchaeology 13, 673e708.

Metcalfe, J.Z., Longstaffe, F.J., Hodgins, G., 2013. Proboscideans and paleoenviron-ments of the Pleistocene Great Lakes: landscape, vegetation, and stable iso-topes. Quat. Sci. Rev. 76, 102e113.

Meltzer, D., 1989. Was stone exchanged among eastern North American Paleo-indians? In: Ellis, C., Lothrop, J. (Eds.), Eastern Paleoindian Lithic Resource Use.Westview Press, Boulder, pp. 11e39.

Meltzer, D., 2004. Modeling the initial colonization of the Americas: issues of scale,demography, and landscape learning. In: Barton, C., Clark, G., Yesner, D.,Pearson, G. (Eds.), The Settlement of the American Continents: a Multidisci-plinary Approach to Human Biogeography. University of Arizona Press, Tucson,pp. 123e137.

Meltzer, D., 2009. First Peoples in a New World: Colonizing Ice Age America. Uni-versity of California Press, Berkeley.

Miller, G.L., 2013. Illuminating activities at Paleo Crossing (33ME274) throughmicrowear analysis. Lithic Technol. 38, 97e108.

Miller, G.L., 2014. Lithic microwear analysis as a means to infer production ofperishable technology: a case from the Great Lakes. J. Archaeol. Sci. 49,292e301.

M.T. Boulanger et al. / Journal of Archaeological Science 53 (2015) 550e558558

Morgan, B., Eren, M.I., Khreisheh, N., Hill, G., Bradley, B., 2014. Clovis bipolar lithicreduction at Paleo Crossing, Ohio: a reinterpretation based on the examinationof experimental replications. In: Jennings, T., Smallwood, A. (Eds.), Clovis: onthe Edge of a New Understanding. Texas A&M Press, College Station (in press).

Morrow, C.A., Elam, J.M., Glascock, M.D., 1992. The use of blue-gray chert in mid-western prehistory. Midcont. J. Archaeol. 17, 166e197.

Mu~niz, M.P., 2014. Determining a cultural affiliation for the CW cache fromNortheastern Colorado. In: Huckell, B., Kilby, D. (Eds.), Clovis Caches: RecentDiscoveries and New Research. University of New Mexico Press, Albuquerque,pp. 107e132.

Nance, J.D., 1984. Lithic exploitation studies in the Lower TennesseeeCumberlandValleys, western Kentucky. In: Butler, B., May, E. (Eds.), Prehistoric ChertExploitation: Studies from the Midcontinent. Center for Archaeological In-vestigations, Southern Illinois University, Carbondale, pp. 101e127.

Nash, D., Coulson, S., Staurset, S., Ullyott, J., Babutsi, M., Hopkinson, L., Smith, M.,2013. Provenancing of silcrete raw materials indicates long distance transportto Tsodilo Hills, Botswana, during the Middle Stone Age. J. Hum. Evol. 64,280e288.

Neff, H., 2000. Neutron activation analysis for provenance determination in archae-ology. In: Ciliberto, E., Spoto, G. (Eds.), Modern Analytical Methods in Art andArchaeology, Chemical Analysis Series, vol. 155. Wiley, New York, pp. 81e134.

Neff, H., 2002. Quantitative techniques for analyzing ceramic compositional data.In: Glowacki, D.M., Neff, H. (Eds.), Ceramic Production and Circulation in theGreater Southwest: Source Determination by INAA and ComplementaryMineralogical Investigations. Cotsen Institute of Archaeology, Los Angeles,pp. 15e36.

O'Brien, M.J., Boulanger, M., Buchanan, B., Collard, M., Lyman, R., Darwent, J., 2014.Innovation and cultural transmission in the American Paleolithic: phylogeneticanalysis of eastern Paleoindian projectile-point classes. J. Anthropol. Archaeol.34, 100e119.

O'Brien, M.J., Lyman, R.L., 2002. The epistemological nature of archaeological units.Anthropol. Theory 2, 37e56.

Pearce, E., 2014. Modelling mechanisms of social network maintenance in hunter-gatherers. J. Archaeol. Sci. 50, 403e413.

Pearce, E., Moutsiou, T., 2014. Using obsidian transfer distances to explore socialnetwork maintenance in late Pleistocene hunter-gatherers. J. Anthropol.Archaeol. 12e20.

Pinson, A., 2011. The Clovis occupation of the Dietz site (35lk1529), Lake County,Oregon, and its bearing on the adaptive diversity of Clovis foragers. Am. Antiq.76, 285e314.

Prasciunas, M., Surovell, T., 2014. Reevaluating the duration of Clovis: the problem ofnon-representative radiocarbon dates. In: Jennings, T., Smallwood, A. (Eds.), Clovis:on the Edge of a New Understanding. Texas A&M Press, College Station (in press).

Redmond, B.G., Tankersley, K., 2005. Evidence of Early Paleoindian bone modifi-cation and use at the Sheriden Cave site (33WY252), Wyandot County, Ohio.Am. Antiq. 70, 503e526.

Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C.,Grootes, P.M., Guilderson, T.P., Haflidason, H., Hajdas, I., Hatt�Z, C., Heaton, T.J.,Hoffmann, D.L., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., Manning, S.W.,Niu, M., Reimer, R.W., Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A.,Turney, C.S.M., van der Plicht, J., 2013. IntCal13 and Marine13 radiocarbon agecalibration curves 0-50,000 years cal BP. Radiocarbon 55 (4).

Sanchez, G., Holliday, V., Gaines, E., Arroyo-Cabrales, J., Martínez-Tagüe~na, N.,Kowler, A., Lange, T., Hodgins, G., Mentzer, S., Sanchez-Morales, I., 2014. Human(Clovis)egomphothere (Cuvieronius sp.) Association ~13,390 Caibrated YBP inSonora, Mexico. Proceedings of the National Academy of Science 111,10972e10977.

Seeman, M., 1994. Intercluster lithic patterning at Nobles Pond: a case for ‘dis-embedded’ procurement among Early Paleoindian societies. Am. Antiq. 59,273e288.

Seeman, M., Loebel, T., Comstock, A., Summers, G., 2013. Working with Wilmsen:Paleoindian end scraper design and use at Nobles Pond. Am. Antiq. 78,407e432.

Shackley, M.S., 2007. Sources of obsidian at the Murray Springs Clovis Site. In:Haynes, C.V., Huckell, B. (Eds.), Murray Springs, a Clovis Site with MultipleActivity Areas in the San Pedro Valley, Arizona, Anthropological Paper No. 71.University of Arizona, Tucson, pp. 250e254.

Sholts, S., Stanford, D., Flores, L., W€arml€ander, S., 2012. Flake scar patterns of Clovispoints analyzed with a new digital morphometrics approach: evidence fordirect transmission of technological knowledge across early North America.J. Archaeol. Sci. 39, 3018e3026.

Smallwood, A., Jennings, T. (Eds.), 2014. Clovis: on the Edge of a New Under-standing. Texas A&M University Press, College Station (in press).

Smith, G.M., Kielhofer, J., 2012. Through the high rock and beyond: placing the LastSupper Cave and Parman Paleoindian lithic assemblages into a regional context.J. Archaeol. Sci. 38, 3568e3576.

Smith, H., Smallwood, A., DeWitt, T., 2014. Defining the normative range of Clovisfluted point shape using geographic models of geometric morphometric vari-ation. In: Jennings, T., Smallwood, A. (Eds.), Clovis: on the Edge of a New Un-derstanding. Texas A&M Press, College Station (in press).

Speer, C., 2014. LA-ICP-MS analysis of Clovis period projectile points from the GaultSite. J. Archaeol. Sci. 52, 1e11.

Speth, J., Newlander, K., White, A., Lemke, A., Anderson, L., 2013. Early Paleoindianbig-game hunting in North America: provisioning or politics? Quat. Int. 285,111e139.

Stanford, D.J., Jodry, M., 1988. The Drake Clovis Cache. Curr. Res. Pleistocene 5,21e22.

Surovell, T., 2000. Early Paleoindian women, children, mobility, and fertility. Am.Antiq. 65, 493e508.

Szabo, J.P., Chanda, A., 2004. Pleistocene glaciation of Ohio, U.S.A. In: Ehlers, J.,Gibbard, P.L. (Eds.), Quaternary Glaciationsdextent and Chronology, Part 2:North America. Elsevier, Amsterdam, pp. 233e236.

Tankersley, K., 1984. Mineraological properties of Wyandotte chert as an aid toarchaeological fingerprinting. In: Vehik, S. (Ed.), Lithic Resource Procurement:Proceedings from the Second Conference on Prehistoric Chert Exploitation,Center for Archaeological Investigations, Occasional Paper No. 4. Southern Illi-nois University.

Tankersley, K., 1989. Late Pleistocene Lithic Exploitation and Human Settlement inthe Midwestern United State. (Ph.D. dissertation). University of Indiana,Bloomington.

Tankersley, K., 2004. The concept of Clovis and the peopling of North America. In:Barton, C., Clark, G., Yesner, D., Pearson, G. (Eds.), The Settlement of theAmerican Continents: a Multidisciplinary Approach to Human Biogeography.University of Arizona Press, Tucson, pp. 49e63.

Tankersley, K., Holland, J., 1994. Lithic procurement patterns at the Paleo Crossingsite, Medina County, Ohio. Curr. Res. Pleistocene 11, 61e63.

Waters, M., Stafford, T., 2007. Redefining the age of Clovis: implications for thepeopling of the Americas. Science 315, 1122e1126.

Waters, M., Stafford, T., Redmond, B., Tankersley, K., 2009. The age of the Paleo-indian assemblage at Sheriden Cave, Ohio. Am. Antiq. 74, 107e111.

Witthoft, J., 1952. A Paleo-Indian site in eastern Pennsylvania: an early huntingculture. Proc. Am. Philos. Soc. 96, 464e495.

Yu, Z., 2000. Ecosystem response to Lateglacial and early Holocene climate oscil-lations in the Great Lakes region of North America. Quat. Sci. Rev. 19, 1723e1747.